Available online at:
https://cc.arcabc.ca/islandora/object/cc%3Apsycjournal

J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.
Submitted Fall 2024; accepted May 2025.

What Determines the Effect of Caffeine Upon Memory?
Authors: Kaydence Kauffman* and Aaliyah Planes
Supervising Instructor: Michael Pollock, Psyc 245 (“Drugs & Behaviour”)
Department of Psychology, Camosun College, 3100 Foul Bay Road, Victoria, BC, Canada V8P 5J2
*Corresponding author email: kaydencekauffman@gmail.com

ABSTRACT
In this paper, we sought to understand what determines whether caffeine affects memory so
that we could learn how to use that to our advantage as students. Previous research has found the
effect of caffeine upon memory is influenced by variables such as sleep, anxiety, and aerobic
exercise. In our correlational study, we tested the strength of these relationships by examining
naturalistic daily changes in their variables longitudinally over a period of two weeks. We
measured caffeine intake by the total mg of caffeine consumed, sleep duration by the total hours
spent sleeping, sleep quality by a subjective scale, anxiety by the State-Trait Anxiety Inventory,
exercise duration by the total number of minutes of aerobic exercise, exercise intensity by a
subjective scale, and memory performance by the n-back task at the end of each day. Data
pooled and standardized across participants in our study showed that memory performance was
not significantly correlated with caffeine intake but was significantly positively correlated with
sleep duration and significantly negatively correlated with both anxiety and exercise intensity.
The relationship between caffeine intake and memory performance was not significantly
moderated by any of the study variables except for sleep duration, with memory performance
showing a positive relationship with caffeine intake when sleep duration was at below average
amounts.
1. Introduction
1.1 Research Problem
Caffeine, a widely known and used
stimulant, has gained significant attention
from college students due to its perceived
influence on cognitive processes, such as
memory. Our research problem regards the
factors that determine whether caffeine
affects our ability to recall and process
information. Our rationale for researching
caffeine’s effects on memory stems from a
desire for academic and personal growth.
Understanding what determines

whethercaffeine impacts memory may help
us optimize its use to enhance daily
cognitive performance, such as improving
focus during work and study. As students,
we want to utilize caffeine to enhance focus,
but also need to understand what factors
could make it could negatively impact
memory, which is essential for active
listening and maintaining relationships.
Gaining insight into this research problem is
important because it could reveal when
caffeine consumption impacts cognitive
functions, potentially helping people make
more informed decisions about their daily
caffeine intake.
82

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

1.2 Literature Review
One factor previously found to influence
the effect of caffeine upon memory is sleep.
For example, in an experimental study by
Onaolapo et al. (2015), adult Swiss Webster
mice of both sexes were assigned to six
groups, including a vehicle control group
and five different doses of caffeine (10, 20,
40, 80, and 120 mg/kg) administered orally
for 14 days. Mice were sleep-deprived for 6
hours pre-training and pre-test. Those in the
sleep deprivation group underwent “gentle
handling” to keep them awake, which
involved tapping on the cage and gently
touching them. The researchers conducted
open-field novel object recognition tests and
Y maze spatial working memory tests on
day 14. The results indicated that a
combination of pre-training and pre-test
sleep deprivation alongside caffeine
impaired both non-spatial and spatial
memory in male and female mice. In the
vehicle control group (non-sleep-deprived
and not given caffeine), memory
performance remained stable across tasks,
serving as a baseline to compare the effects
of caffeine at different doses. Non-sleepdeprived mice that received caffeine
exhibited dose-dependent effects: lower
doses (such as 10 and 20 mg/kg) led to
improved performance in the novel object
recognition test compared to the vehicle
control group, while higher doses (such as
80 and 120 mg/kg) resulted in impairments
in both spatial memory tasks and the Y maze
task. In contrast, in sleep-deprived mice,
memory performance declined across all
caffeine doses, with higher doses causing
more significant impairments in both spatial
and non-spatial memory tasks. Based on
these results, the researchers suggested that
sleep deprivation may alter the cognitive
effects of caffeine, adversely impacting
memory performance.

Another factor previously found to
influence the effect of caffeine upon
memory is anxiety levels. For example, in an
experimental study by Terry and Phifer
(1986), the independent variable was the
consumption of caffeine. The researchers
manipulated this variable by giving
participants either 100 mg of caffeine
dissolved in Gatorade or a placebo
(Gatorade with a small amount of table salt).
This was done through a double-blind
procedure, meaning neither the participants
nor the researchers knew who was receiving
caffeine or the placebo, which helped reduce
bias. The dependent variable was memory
performance, measured by the participant's
ability to recall words from a list. This was
assessed using the Auditory Verbal Learning
Test (AVLT), where participants were asked
to remember and recall a series of words
(common nouns) across multiple trials, both
immediately and after a delay. The
researchers used the number of words
recalled to assess memory performance. The
results showed that participants who
consumed caffeine recalled fewer words
compared to those in the control group. The
caffeine group particularly struggled with
recalling words from the middle-to-end
portions of the lists, indicating a negative
effect of caffeine on memory retention.
Based on these results, the researchers
suggested that caffeine, contrary to common
belief, may impair memory performance,
especially when recalling newly learned
information. Anxiety was measured using
the neuroticism scale from the Maudsley
Personality Inventory and while higher
anxiety levels were linked to better recall in
the pretest, the combination of anxiety and
caffeine worsened memory performance
overall. A third factor previously found to
influence the effect of caffeine upon
memory is aerobic exercise. For example, in
an experimental study by Morava et al.
83

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

(2019), the researchers recruited twenty-nine
non-caffeine consumers and thirty caffeine
consumers aged 18-40 and utilized a
randomized counterbalanced crossover
design across two phases. Participants
completed the n-back task to assess working
memory after receiving either caffeine or
aerobic exercise interventions. The exercise
intervention consisted of a single session of
moderate-intensity aerobic exercise on a
Woodway PPS treadmill, which included a
2-5 minute warm-up walk, 15 minutes of
walking at moderate intensity (defined as 40
to 60% of Heart Rate Reserve), and a 2-5
minute cool-down walk. The caffeine
administration intervention consisted of oral
administration of powdered caffeine. Each
participant was given 1.2 mg/kg (body
weight) of powdered caffeine dissolved in
100 mL of water. The findings revealed
significant improvements in working
memory accuracy for caffeine consumers
after both caffeine and exercise conditions
compared to baseline. Notably, when
caffeine and exercise were combined, there
were even greater enhancements in accuracy
and faster reaction times on the n-back test,
suggesting a synergistic effect between the
two interventions. Additionally, caffeine
withdrawal symptoms, including memory
impairments measured through the n-back
task, were significantly reduced following
caffeine administration after a 12-hour
deprivation period. However, this effect was
not influenced by anxiety levels or exercise,
as neither showed significant interactions
with caffeine intake on memory
performance. Based on these results, the
researchers suggested that in regular
consumers caffeine may enhance working
memory performance and alleviate memory
impairments caused by caffeine withdrawal.

1.3 Hypotheses
Based on the above literature review, we
predicted the following hypotheses:
● If sleep decreases then increasing
caffeine intake will decrease memory
performance.
● If anxiety increases then increasing
caffeine intake will decrease memory
performance.
● If aerobic exercise increases then
increasing caffeine intake will increase
memory performance.
2. Methods
2.1 Participants
The two authors of this paper served as
the participants in this study. They were
both female undergraduate students at
Camosun College, completing this study as
an assignment for PSYC 245: Drugs &
Behaviour Psychology. Both participants
were regular caffeine consumers and were
grouped together based on a mutual interest
in studying caffeine’s effects on memory.
Their ages ranged from 18-20 years old,
with an average age of 19. The first
participant had been consuming caffeinated
beverages (including coffee, carbonated
beverages, and tea) regularly for
approximately six years, indicating moderate
tolerance as she could drink multiple cups of
caffeine without feeling significant effects.
The second participant had been a regular
consumer of caffeine (primarily tea and
coffee) for about five years, also selfreporting moderate tolerance due to her
ability to consume similar amounts without
noticeable effects. Both participants met the
study’s inclusion criteria of regular caffeine
consumption, which was necessary to assess
the potential effects on memory performance
84

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

found in regular caffeine users by Morava et
al. (2019).
2.2 Materials and Procedures
We performed a correlational study to
test concurrently all of our hypotheses by
examining naturalistic daily changes in their
variables longitudinally. Each participant
kept a study journal with them at all times
over this study’s one-week period in order to
record self-observations of the following
five variables: (1) caffeine intake, (2) sleep,
(3) anxiety, (4) aerobic exercise, and (5)
memory performance.
2.2.1 Caffeine Intake
Over the course of 14 days, the
participants implemented a detailed tracking
method to capture day-to-day variations in
caffeine intake. Each participant utilized a
digital logging application called Barista to
accurately record daily caffeine
consumption from various sources,
including coffee, tea, energy drinks, and
caffeinated snacks (see Appendix A for the
sample Barista log format). Each entry
specified the type and amount consumed,
measured in millilitres, grams, and
milligrams of caffeine, allowing for precise
quantification. This systematic approach
provided valuable insights into individual
caffeine habits.
2.2.2 Sleep
Over the course of 14 days, participants
implemented a detailed tracking method to
capture daily variations in sleep. Participants
used a dedicated sleep diary to record both
objective and subjective sleep data (see
Appendix B for the sleep diary template).
Each participant documented the time of
falling asleep and the time of waking up
each night, allowing for precise calculation
of total sleep hours. Additionally, each

morning, participants rated sleep quality
using a five-point scale from one (very poor)
to five (excellent). This combination of
objective timing and subjective quality
ratings provided a comprehensive view of
sleep patterns, with numerical scores for
both sleep duration and quality captured for
each previous night of the study and
statistically compared with the values of the
other study variables collected during the
daytime immediately following each of
those nights.
2.2.3 Anxiety
To measure anxiety levels in participants,
the State-Trait Anxiety Inventory (STAI)
was utilized, available online at
ToolOnline.net (see Appendix C for STAI
scoring guide). The STAI consisted of 40
items divided into two scales: the State
Anxiety scale, which assessed how
participants felt at the moment, and the Trait
Anxiety scale, which evaluated their general
anxiety levels over time. Each participant
completed the State Anxiety scale each
morning for 14 days, rating their feelings on
a four-point Likert scale from one (not at all
anxious) to four (very anxious; see
Appendix C for details on the Likert scale
and scoring guide). This daily assessment
captured fluctuations in anxiety levels,
allowing for a sensitive measure of how
anxiety might impact memory performance.
Participants also completed the Trait
Anxiety scale once at the beginning of the
study to provide a baseline measure of their
general anxiety levels. The online platform
automatically scored the inventory and
provided feedback on each participant's
anxiety levels. This comprehensive tracking
method ensured sensitivity to daily
variations.

85

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

2.2.4 Aerobic Exercise
Over the course of 14 days, the
participants implemented a detailed tracking
method to capture day-to-day variations in
aerobic exercise levels. Each participant
logged all aerobic exercise activities in a
daily exercise journal (see Appendix D for
exercise journal template). Each entry
detailed the type of exercise performed (e.g.,
running, cycling, swimming), along with the
duration (measured in minutes) and
perceived intensity, rated on a scale of one
(very light) to five (maximum effort). This
approach enabled tracking of both exercise
frequency and intensity over the study
period, providing insight into potential
variations in aerobic activity levels. Each
day’s overall exercise level was calculated
by determining the total duration of exercise
completed and averaging the intensity of
those exercises. On days when no exercise
occurred, intensity values were not recorded.
2.2.5 Memory Performance
To measure memory performance, each
participant completed a daily n-back task
using the Brain Workshop software (see
Appendix E for n-back task setup and
instructions). The task, administered on a
computer, was designed to assess working
memory by prompting participants to recall
sequences of visual or auditory stimuli
presented in intervals (e.g., two-back or
three-back). Each session lasted
approximately five minutes and
automatically recorded response accuracy
(percentage of correct responses) and
response time for each trial. Data was saved
in a digital format, capturing both response
accuracy and latency, ensuring detailed
tracking of memory outcomes. Participants
completed the task daily for a minimum of
14 days, typically in the evening after
recording that day’s other variable
measurements. This timing allowed for

potential influences from those variables,
such as caffeine intake, anxiety levels, and
exercise, to be reflected in the memory
performance data.
2.3 Planned Analyses
In this study, statistical analyses were
performed using Microsoft Excel (version
2024) to evaluate the data. To assess the
strength and statistical significance of
associations that memory has with caffeine
intake, sleep, anxiety, and exercise, we
performed Pearson product-moment
correlations of accuracy scores on the nback task with mg of caffeine consumed,
hours of sleep attained, subjective sleep
quality scores, STAI scores, minutes of
exercise attained, and subjective exercise
intensity scores. We performed all of the
above correlations separately for each
participant as well as using data pooled
across all of the participants. To assess the
moderating influence of the other variables
predicted by our three hypotheses, we
performed multiple regression analyses of
their predictor variable (caffeine intake),
their potential moderating variable (sleep,
anxiety, or aerobic exercise), an interaction
term (creating by multiplying the values of
the predictor variable and the potential
moderating variable for each day), and their
outcome variable (memory performance).
For testing Hypothesis #1 that sleep duration
moderates the relationship between caffeine
and memory, we examined the regression
output to see if the interaction term created
with values of mg of caffeine consumed
multiplied by either hours of sleep attained
or subjective sleep quality scores was
statistically significant. For testing
Hypothesis #2 that anxiety moderates the
relationship between caffeine and memory,
we examined the regression output to see if
the interaction term created with values of
86

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

mg of caffeine consumed multiplied by
STAI scores was statistically significant. For
testing Hypothesis #3 that exercise
moderates the relationship between caffeine
and memory, we examined the regression
output to see if the interaction term created
with values of mg of caffeine consumed
multiplied by either minutes of exercise
attained or subjective exercise intensity
scores was statistically significant. For the
correlation and regression analyses using
pooled data, in addition to using the raw
data, we also performed analyses after we
had first transformed the data from each
participant into z-scores in order to
standardize differences in averages and
variability seen between the participants in
their data and thus make them more
comparable. A coefficient was considered
statistically significant if the probability of
its random occurrence (p) was < .05 (i.e.,
less than 5% of the time expected by chance
alone).
3. Results
3.1 Correlational Analyses
The correlational analyses revealed no
significant relationships between memory
performance and caffeine intake but did find
memory performance was significantly
correlated with measures of sleep, anxiety,
and exercise (see Table 1). Caffeine intake
was not significantly correlated with
memory performance in either of the two
participants’ data (r = -.36 & .42, p = .21 &
.14), the pooled raw data (r = .15, p = .46,
see Figure 1), or the pooled standardized
data (r = .03, p = .87, see Figure 2). Sleep
duration was significantly correlated with
memory performance in one of the two
participants’ data (r = .13 & .78, p = .68 &
.0005), the pooled raw data (r = .47, p = .01,
see Figure 3), and the pooled standardized

data (r = .45, p = .01, see Figure 4). Sleep
quality was significantly correlated with
memory performance in one of the two
participants’ data (r = -.16 & .90, p = .59 &
.0000007) and the pooled raw data (r = .38,
p = .04, see Figure 5), but not in the pooled
standardized data (r = .37, p = .05, see
Figure 6). Anxiety was significantly
correlated with memory performance in both
of the two participants’ data (r = -.57 & -.70,
p = .03 & .004), the pooled raw data (r = 0.57, p = .001, see Figure 7), and the pooled
standardized data (r = -.64, p = .0002, see
Figure 8). Exercise duration was not
significantly correlated with memory
performance in either of the two
participants’ data (r = .12 & .24, p = .70 &
.41), the pooled raw data (r = .10, p = .60,
see Figure 9), or the pooled standardized
data (r = .18, p = .36, see Figure 10).
Exercise intensity was not significantly
correlated with memory performance in
either of the two participants’ data (r = -.47
& -.35, p = .09 & .30) or the pooled raw data
(r = -.37, p = .07, see Figure 11), but was
significantly correlated in the pooled
standardized data (r = -.42, p = .04, see
Figure 12). Since no exercise was performed
on three of the study days for one of the
participants, the above correlations of
exercise intensity included a total of just 25
days of pooled data instead of the full 28
days (four-weeks) of pooled data included in
the other correlations. In a comparison of all
the correlation coefficients when using the
pooled standardized data for their analyses,
the variable that showed the strongest
correlation with memory performance was
anxiety with an r-value of -.64, indicating a
large effect size (r ≥ 0.50).
3.2 Multiple Regression Analyses
The multiple regression analyses revealed
that the relationship between caffeine intake
87

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

and memory performance was not
significantly moderated by any of the study
variables except for sleep duration (see
Table 2). The interaction of caffeine with
sleep duration did not significantly predict
memory performance in the pooled raw data
(t = -.22, p = .83; see Figure 13) but did in
the pooled standardized data (t = -2.10, p =
.046; see Figure 14). Visual inspection of the
pooled standardized data in Figure 14
showed that memory performance displayed
a positive relationship with caffeine intake
when sleep duration was at levels below
average (sleep duration z-scores < 0), close
to no relationship with caffeine intake when
sleep duration was at low-to-moderate levels
above average (sleep duration z-scores = 0
to 1), and a negative relationship with
caffeine intake when sleep duration was at
high levels above average (sleep duration zscores > 1). Conversely, by treating sleep
duration as the predictor variable and
caffeine intake as the moderator variable
(see Figure 15), this same pooled
standardized data can be interpreted as
showing that memory performance
displayed a positive relationship with sleep
duration when caffeine intake was at levels
below average amounts (caffeine intake zscores < 0), close to no relationship with
sleep duration when caffeine intake was at
low-to-moderate levels above average
(caffeine intake z-scores = 0 to 0.75), and a
negative relationship with sleep duration
when caffeine intake was at high levels
above average (caffeine intake z-scores >
0.75). The significant interaction between
caffeine intake and sleep duration upon
memory performance occurred despite no
significant correlation found directly
between sleep duration and caffeine intake
in either of the two participants’ data (r = .39 & .25, p = .17 & .40), the pooled raw
data (r = .22, p = .26, see Figure 16), or the
pooled standardized data (r = -.07, p = .73,

see Figure 17). In contrast to the significant
interaction of caffeine intake with sleep
duration, the interaction of caffeine intake
with all of the other study variables did not
significantly predict memory performance in
either the pooled raw data (all absolute t ≤
.94, all p ≥ .36) or the pooled standardized
data (all absolute t ≤ 1.96, all p ≥ .06).
4. Discussion
4.1 Summary of Results
Based on previous research, we
hypothesized that changes in three variables
would moderate the relationship between
caffeine and memory performance: lower
sleep durations would predict increases in
caffeine intake being associated with
decreases in memory performance
(Hypothesis #1), higher anxiety levels would
predict increases in caffeine intake being
associated with decreases in memory
performance (Hypothesis #2), and higher
levels of aerobic exercise would predict
increases in caffeine intake being associated
with increases in memory performance
(Hypothesis #3). While data pooled and
standardized across participants in our study
did show that caffeine intake had a
significant interaction with sleep duration (p
< 0.05) and a trend towards a significant
interaction with sleep quality (p = 0.06)
upon memory performance, these findings
were in the opposite direction from what we
predicted, with improved instead of
impaired memory performance being
associated with above average caffeine
intake levels when sleep durations were
below average. Anxiety and exercise failed
to show any significant interactions with
caffeine intake upon memory performance.
These results indicate that none of our
hypotheses were supported by our data.
88

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

4.2 Relation of Results to Past Research
The findings in our study, which revealed
that when sleep duration is low then
increased caffeine intake is associated with
increased memory performance, differed
from previous research done by Onaolapo et
al. (2015). Their research demonstrated
significant impairments in memory due to
the combined effects of sleep deprivation
and caffeine in mice (Onaolapo et al., 2015).
Onaolapo et al. (2015) employed an
experimental research design to detect the
detrimental effects of caffeine. While our
correlational research design was not able to
similarly establish whether any relationships
between variables were causal ones, it
should still have been sufficiently sensitive
to detect the existence and direction of any
relationships in general, independent of
whether they are causal or non-causal ones.
In our research, the pattern of interaction
found between sleep duration and caffeine
intake upon memory performance suggests
that unlike the results of experiments done
with mice, increased caffeine intake when
sleep is disturbed does not impair memory
performance in human participants under
naturalistic conditions. The methodological
variations —such as species differences,
measurement tools, and environmental
contexts—likely affected the difference in
our findings compared to past research
(Onaolapo et al., 2015). To test whether
increased caffeine intake actually causes
enhanced memory performance when sleep
duration was low, future studies should
adopt an experimental design in human
participants. Additionally, longitudinal
designs could explore whether chronic
caffeine use interacts with habitual sleep
patterns to impact memory differently from
the acute conditions explored in the current
study. These refinements could clarify the

conditions under which caffeine and sleep
disturbances affect memory.
Our findings differed from those of Terry
and Phifer (1986) who observed a negative
relationship between anxiety levels and
caffeine's impact on memory performance.
In their experimental study, higher anxiety
levels combined with caffeine consumption
were associated with worsened memory
recall, particularly for information learned
later in a sequence (Terry & Phifer, 1986).
However, in our study, while we observed a
large negative correlation between anxiety
and memory performance, caffeine intake
did not significantly moderate this
relationship or show a significant correlation
directly with memory performance. Our
study measured anxiety using the State-Trait
Anxiety Inventory (STAI), a standardized
and widely used tool, similar to the
neuroticism scale used in the Terry and
Phifer (1986) study. The lack of significant
findings in our study may be attributed to
our reliance on naturalistic caffeine
consumption, which lacked the controlled
dose manipulation present in Terry and
Phifer’s work (Terry & Phifer, 1986). In
their study, participants received
standardized caffeine doses, ensuring
consistent intake levels across individuals.
In contrast, our participants consumed
varying amounts of caffeine based on their
typical daily habits, which may have
introduced inconsistencies in dosage. These
variations could have reduced the likelihood
of detecting significant effects, particularly
if some participants consumed doses too low
to influence memory performance.
Additionally, the range of caffeine doses in
our study may have differed from those
tested by Terry and Phifer, further
contributing to the lack of significant results.
Additionally, while we used the n-back task
to measure working memory performance,
Terry and Phifer employed the Auditory
89

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Verbal Learning Test (AVLT), which
focuses on verbal recall (Terry & Phifer,
1986). These methodological differences
might explain the discrepancies between our
findings and theirs. Future research should
combine the controlled dosing approach
with diverse memory assessments to better
explore the interaction between anxiety,
caffeine, and various types of memory
performance. The findings in our research
which, although showing a significant
correlation between exercise intensity and
memory performance in the pooled
standardized data, failed to show a
significant interaction between caffeine
intake and exercise upon memory
performance, contrast with earlier research
done by Morava et al. (2019). Their study
demonstrated synergistic enhancements in
working memory accuracy and reaction
times when caffeine consumption was
combined with aerobic exercise. Morava et
al. (2019) observed significant
improvements in cognitive performance,
such as faster reaction times and greater
accuracy on the n-back test, under controlled
conditions. The differences in findings
might stem from methodological variations
between the two studies. Terry and Phifer
(1986) conducted a controlled experimental
study, systematically manipulating caffeine
and exercise variables and focusing on
immediate cognitive outcomes. This
controlled design ensured precise dosage
administration and timing, reducing
variability and enhancing the detection of
potential synergistic effects (Morava et al.,
2019). In contrast, our observational study
relied on self-reported data for caffeine
intake and exercise, which introduces
variability and potential inaccuracies.
Participants may have misestimated their
caffeine intake, particularly if they
consumed beverages with unclear labeling
or variable caffeine content. Similarly, self-

reported exercise data may have been
imprecise, especially regarding intensity
levels. These inaccuracies, combined with
natural fluctuations in participants’ daily
habits, may have reduced our study’s
sensitivity to detect subtle or complex
interactions. Furthermore, the Morava et al.
(2019) investigation included participants
experiencing withdrawal symptoms, which
may have amplified the observed effects of
caffeine. This context-dependent influence
was not accounted for in our study,
potentially contributing to the discrepancy.
Future studies could explore the role of
withdrawal symptoms, caffeine tolerance,
and the timing of exercise relative to
caffeine consumption to better understand
the context-dependent effects. Longitudinal
studies could also examine whether habitual
patterns of exercise and caffeine intake
interact over time to influence memory
performance.
4.3 Implications of Results
Our findings suggest that anxiety and
exercise did not significantly influence the
relationship between caffeine intake and
memory performance. This indicates that
students seeking to optimize their cognitive
performance through caffeine use may not
need to rely on strategies like exercise to
counteract potential negative effects.
Instead, maintaining consistent caffeine
consumption habits and being mindful of
individual responses may be more effective.
These insights can help students make
informed decisions about their caffeine
intake when managing academic demands,
focus, and cognitive well-being.
Our research aimed to explore the factors
that determine whether caffeine impacts
memory performance, particularly the roles
of sleep, anxiety, and exercise. We hoped
this knowledge would provide practical
90

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

guidance for students looking to enhance
focus and cognitive performance. While our
findings did not support our hypotheses,
they suggest that caffeine’s effects on
memory may be less dependent on anxiety
and exercise than previously thought. This
outcome emphasizes the need for further
research to better understand the conditions
under which caffeine consumption
influences memory.
References
Morava, A., Fagan, M. J., & Prapavessis, H.
(2019). Effects of caffeine and acute

aerobic exercise on working memory and
caffeine withdrawal. Scientific Reports,
9(1), 19644. doi:10.1038/s41598-01956251-y
Onaolapo, O. J., Onaolapo, A. Y., Akanmu,
M. A., & Olayiwola, G. (2015).
Caffeine/sleep- deprivation in mice
produces complex memory effects.
Annals of Neurosciences, 22(3), 139-149.
doi : 10.5214/ans.0972.7531.220304
Terry, W. S., & Phifer, B. (1986). Caffeine
and memory performance on the AVLT.
Journal of Clinical Psychology, 42(6),
860–863. doi:10.1002/10974679(198611)42:6<860::AIDJCLP2270420604>3.0.CO;2-T

91

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Table 1
Correlational Analyses of Study Variables

Study Variables

Participant
#1

Participant
#2

Pooled raw
data

Pooled
standardize
d data

r

n

r

n

r

n

r

n

Caffeine intake & memory performance

-0.36

14

0.42

14

0.15

28

0.03

28

Sleep duration & memory performance

0.13

14

0.78*

14

0.47*

28

0.45*

28

Sleep quality & memory performance

-0.16

14

0.90*

14

0.38*

28

0.37

28

Anxiety & memory performance

-0.57*

14

-0.70*

14

-0.57*

28

-0.64*

28

Exercise duration & memory performance

0.12

14

0.24

14

0.10

28

0.18

28

Exercise intensity & memory performance

-0.47

14

-0.35

11

-0.37

25

-0.42*

25

* p < .05.

92

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Table 2
Multiple Regression Analyses of Interaction Terms for Caffeine Intake with Study Variables on
Memory Performance Using Either Raw or Standardized (z-Scores) Pooled Data
Interaction Term

df

t-Statistic

p-Value

Raw

z-Score

Raw

zScore

Raw

zScore

Caffeine intake * sleep duration

24

24

-0.22

-2.10

0.83

0.046

Caffeine intake * sleep quality

24

24

0.11

-1.96

0.91

0.06

Caffeine intake * anxiety

24

24

-0.36

0.17

0.72

0.87

24

24

-0.94

-1.10

0.36

0.28

21

21

-0.55

-0.90

0.59

0.38

Caffeine intake * exercise
duration
Caffeine intake * exercise
intensity

Note. df = the residual degrees of freedom which was calculated by the total number of
observations (25 or 28 days) minus the number of predictors (3 including the interaction term)
minus one, t-statistic = the ratio of the difference in a number's known value from its expected
value over the standard error of the distribution it was derived from, p-value = the probability
(expressed as a proportion) that a given set of data occurred by chance alone, and z-score = the
number of standard deviations away that a value is from the mean of the distribution it was
derived from. p < .05 indicates statistical significance.

93

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Figure 1
Association Between Caffeine Intake and Memory Performance Using Pooled Raw Data

Memory performance (% of correct
responses)

120
100

80
60
40
20
0
0

100

200

300
Caffeine (mg)

400

500

600

Notes. Marker colour differentiates participants: red = participant #1 and orange = participant #2.
Some data might not be visible in the figure due to overlapping markers.

94

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Figure 2
Association Between Caffeine Intake and Memory Performance Using Pooled Standardized Data
2

Memory performance (z-score)

1.5

-3

1
0.5
0
-2

-1

-0.5

0

1

2

3

-1
-1.5
-2
-2.5
Caffeine (z-score)

Notes. Marker colour differentiates participants: red = participant #1 and orange = participant #2.
Some data might not be visible in the figure due to overlapping markers.

95

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Figure 3
Association Between Sleep Duration and Memory Performance Using Pooled Raw Data

Memory performance (% of correct
responses)

120
100

80
60
40
20
0
0

2

4

6
8
Sleep duration (hours)

10

12

14

Notes. Marker colour differentiates participants: red = participant #1 and orange = participant #2.
Some data might not be visible in the figure due to overlapping markers.

96

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Figure 4
Association Between Sleep Duration and Memory Performance Using Pooled Standardized Data
2

Memory performance (z-score)

1.5

-3

1
0.5
0
-2.5

-2

-1.5

-1

-0.5

-0.5

0

0.5

1

1.5

2

-1
-1.5
-2
-2.5
Sleep duration (z-score)

Notes. Marker colour differentiates participants: red = participant #1 orange = participant #2.
Some data might not be visible in the figure due to overlapping markers.

97

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Figure 5
Association Between Sleep Quality and Memory Performance Using Pooled Raw Data

Memory performance (% of correct
responses)

120
100

80
60
40
20
0
0

1

2
3
4
5
Sleep quality (1-5 scale: 1 = very poor; 5 = excellent)

6

Notes. Marker colour differentiates participants: red = participant #1 and orange = participant #2.
Some data might not be visible in the figure due to overlapping markers.

98

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Figure 6
Association Between Sleep Quality and Memory Performance Using Pooled Standardized Data
2

Memory performance (z-score)

1.5

-3

1
0.5
0
-2.5

-2

-1.5

-1

-0.5

-0.5

0

0.5

1

1.5

2

-1
-1.5
-2
-2.5
Sleep quality (z-score)

Notes. Marker colour differentiates participants: red = participant #1 and orange = participant #2.
Some data might not be visible in the figure due to overlapping markers.

99

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Figure 7
Association Between Anxiety and Memory Performance Using Pooled Raw Data

Memory performance (% of correct
responses)

120
100

80
60
40
20
0
0

10

20
30
40
50
60
70
Anxiety (state anxiety score possibly ranging from 20-80)

80

Notes. Marker colour differentiates participants: red = participant #1 and orange = participant #2.
Some data might not be visible in the figure due to overlapping markers.

100

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Figure 8
Association Between Anxiety and Memory Performance Using Pooled Standardized Data
2

Memory performance (z-score)

1.5

-2

1
0.5
0
-1.5

-1

-0.5

-0.5

0

0.5

1

1.5

2

2.5

-1
-1.5
-2
-2.5
Anxiety (z-score)

Notes. Marker colour differentiates participants: red = participant #1 and orange = participant #2.
Some data might not be visible in the figure due to overlapping markers.

101

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Figure 9
Association Between Exercise Duration and Memory Performance Using Pooled Raw Data

Memory performance (% of correct
responses)

120
100

80
60
40
20
0
0

100

200

300
400
500
Exercise duration (minutes)

600

700

Notes. Marker colour differentiates participants: red = participant #1 and orange = participant #2.
Some data might not be visible in the figure due to overlapping markers.

102

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Figure 10
Association Between Exercise Duration and Memory Performance Using Pooled Standardized
Data
2

Memory performance (z-score)

1.5

-1.5

1
0.5

0
-1

-0.5

-0.5

0

0.5

1

1.5

2

2.5

3

-1
-1.5
-2
-2.5

Exercise duration (z-score)

Notes. Marker colour differentiates participants: red = participant #1 and orange = participant #2.
Some data might not be visible in the figure due to overlapping markers.

103

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Figure 11
Association Between Exercise Intensity and Memory Performance Using Pooled Raw Data

Memory performance (% of correct
responses)

120
100

80
60
40
20
0
0

0.5
1
1.5
2
2.5
3
3.5
4
Exercise intensity (1-5 scale: 1 = very light; 5 = maximum effort)

4.5

Notes. Marker colour differentiates participants: red = participant #1 and orange = participant #2.
Some data might not be visible in the figure due to overlapping markers.

104

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Figure 12
Association Between Exercise Intensity and Memory Performance Using Pooled Standardized
Data
2

Memory performance (z-score)

1.5

-1.5

1
0.5

0
-1

-0.5

-0.5

0

0.5

1

1.5

2

2.5

-1
-1.5
-2
-2.5

Exercise intensity (z-score)

Notes. Marker colour differentiates participants: red = participant #1 and orange = participant #2.
Some data might not be visible in the figure due to overlapping markers.

105

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Figure 13
Association Between Sleep Duration and Memory Performance for Different Caffeine Intake
Levels Using Pooled Raw Data

Memory performance (% of correct responses)

120

100

0 to <100

80

>100 to 150
>150 to 200

60

>200 to 555
Linear (0 to <100)

40

Linear (>100 to 150)
Linear (>150 to 200)
Linear (>200 to 555)

20

0
0

2

4

6
8
10
Sleep duration (hours)

12

14

Notes. Marker colour differentiates caffeine intake: red = 0 to <100 mg, orange = >100 to 150
mg, yellow = >150 to 200 mg, and green = >200 to 555 mg. Some data might not be visible in
the figure due to overlapping markers.

106

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Figure 14
Association Between Sleep Duration and Memory Performance for Different Caffeine Intake
Levels Using Pooled Standardized Data
2
1.5

Memory performance (z-score)

1

-3

-2.0 to -0.75

0.5

-0.75 to 0
0 to 0.75

0
-2

-1

0
-0.5

1

2

0.75 to 3.0
Linear (-2.0 to -0.75)
Linear (-0.75 to 0)

-1
-1.5

Linear (0 to 0.75)
Linear (0.75 to 3.0)

-2
-2.5
Sleep duration (z-score)

Notes. Marker colour differentiates caffeine intake: red = -2.0 to -0.75 z, orange = -0.75 to 0 z,
yellow = 0 to 0.75 z, and green = 0.75 to 3.0 z. Some data might not be visible in the figure due
to overlapping markers.

107

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Figure 15
Association Between Caffeine Intake and Memory Performance for Different Sleep Duration
Amounts Using Pooled Standardized Data
2
1.5
-2.5 to -1.5

Memory performance (z-score)

1

-1.5 to -0.5
-0.5 to 0

0.5

0 to 0.5
0

0.5 to 1.0
1.0 to 2.0

-0.5

Linear (-2.5 to -1.5)
-1

Linear (-1.5 to -0.5)
Linear (-0.5 to 0)

-1.5

Linear (0 to 0.5)
Linear (0.5 to 1.0)

-2

Linear (1.0 to 2.0)
-2.5
-3

-2

-1
0
1
Caffeine intake (z-score)

2

3

Notes. Marker colour differentiates sleep duration: dark red = -2.5 to -1.5 z, light red = -1.5 to 0.5 z, orange = -0.5 to 0 z, yellow = 0 to 0.5 z, light green = 0.5 to 1.0 z, and dark green = 1.0 to
2.0 z. Some data might not be visible in the figure due to overlapping markers.

108

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Figure 16
Association Between Sleep Duration and Caffeine Intake Using Pooled Raw Data
600

Caffeine (mg)

500

400
300
200
100
0
0

2

4

6
8
Sleep duration (hours)

10

12

14

Notes. Marker colour differentiates participants: red = participant #1 and orange = participant #2.
Some data might not be visible in the figure due to overlapping markers.

109

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Figure 17
Association Between Sleep Duration and Caffeine Intake Using Pooled Standardized Data
3

Caffeine (z-score)

2

-3

1
0
-2.5

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

-1
-2
-3
Sleep duration (z-score)

Notes. Marker colour differentiates participants: red = participant #1 and orange = participant #2.
Some data might not be visible in the figure due to overlapping markers.

110

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

Appendix A: Sample Barista Log Format for Caffeine Intake Tracking

Date

Time

Beverage/
Food Item

Source

Amount
(mL or
grams)

Caffeine
Content
(mg)

Notes

2024-10-01 8:00 AM

Coffee

Brewed,
dark roast

240 mL

95 mg

Home-brewed

2024-10-01 11:30 AM

Tea

Black tea

200 mL

47 mg

Cafepurchased

2024-10-01 2:00 PM

Energy
drink

Red Bull

250 mL

80 mg

Store-bought

2024-10-01 5:15 PM

Chocolate
bar

Dark
chocolate

50 g

20 mg

Estimated from
CNF
(Canadian
Nutrient File)

2024-10-01

Daily Total

-

-

245 mg

-

Note: Participants should note if caffeine content was estimated from the Canadian Nutrient File
(CNF) or other nutritional information sources.

Appendix B: Sleep Diary Template
Participant Name: ______________________
Date: ______________________
Sleep Entry
1. Time went to bed: ______________________
2. Time fell asleep: ______________________
3. Time woke up: ______________________
4. Total sleep duration (in hours): ______________________

111

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

5. Sleep quality rating (1-5): ______________________
(1 = Very Poor, 2 = Poor, 3 = Fair, 4 = Good, 5 = Excellent)
6. Notes/Comments:
○
○

Reminder: Please record times consistently in either 24-hour format or AM/PM format.
Appendix C: Scoring Guide for State-Trait Anxiety Inventory (STAI)

Scale

Item

Description

Numbers
State

Items

Anxiety

1-20

Trait

Items

Anxiety

21-40

Score

Interpretation
Guide

Range
Measures current 20-80
feelings of
anxiety or
calmness,
reflecting the
participant's
anxiety “state” at
a given moment.

Scores are
categorized as
low (20-39),
moderate (4059), or high (6080) levels of
state anxiety.

Assesses general
and enduring
anxiety
characteristics,
indicating a
baseline level of
anxiety "trait."

Scores are
categorized as
low (20-39),
moderate (4059), or high (6080) levels of trait
anxiety.

20-80

Scoring and Likert Scale Details:

112

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

● Likert Scale for STAI: Each item is rated on a four-point Likert scale as follows:
○ 1 - Not at all anxious
○ 2 - Somewhat anxious
○ 3 - Moderately anxious
○ 4 - Very anxious
Note: Lower scores indicate less anxiety, while higher scores indicate greater anxiety for both
State and Trait scales.
● Daily State Anxiety Scores: Participants completed the State Anxiety scale each morning
for seven days to capture day-to-day fluctuations in anxiety.
● Trait Anxiety Baseline Score: The Trait Anxiety scale was administered once at the
beginning of the study to establish a baseline of each participant’s general anxiety.
● Automatic Scoring: ToolOnline.net automatically scored each inventory and categorized
scores into low, moderate, or high anxiety levels based on the participant’s responses.

Appendix D: Exercise Journal Template
Participant Name: ______________________
Date: ______________________
Exercise Entry
1. Type of exercise:
○ Running
○ Cycling
○ Swimming
○ Walking
113

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

○ Other:
2. Duration of Exercise (in minutes): ______________________
3. Perceived intensity rating (1-5):
(1 = Very Light, 2= Light, 3 = Moderate, 4 = Hard, 5 = Maximum Effort)
Note: If none of the provided intensity ratings feel accurate, please indicate this in the
notes/comments.
4. Notes/Comments:
○
○

Appendix E: Setup and Instructions for N-Back Task
Task Overview: The n-back task was administered using Brain Workshop software, a tool
designed to evaluate working memory by requiring participants to recall sequences of stimuli
presented at specified intervals. The task involved two variations: visual and auditory n-back
sequences.
Setup Instructions
1. Software installation: Brain Workshop software was installed on each participant’s
computer.
2. Task configuration:
○ Interval level: Each session was set to a two-back or three-back configuration,
based on the study’s design.
114

Planes & Kauffman - J Camosun Psyc Res. (2025). Vol. 7(1), pp. 82-115.

○ Stimulus type: Participants completed either a visual n-back (identifying the
position of visual stimuli) or an auditory n-back (recognizing a sound sequence).
○ Session length: Each session was configured to last approximately five minutes,
automatically transitioning through trials.
3. Quiet Environment: It is recommended that participants complete this task in a quiet,
distraction-free environment to ensure accuracy.
Participant Instructions:
1. Starting the task: Participants opened the Brain Workshop software and selected the
assigned n-back level for that session (two-back or three-back).
2. Responding to stimuli:
○ Visual task: Participants pressed a designated key if the current visual stimulus
matched the one shown "n" steps earlier.
○ Auditory task: Participants pressed a designated key if the current auditory
stimulus matched the sound heard "n" steps earlier.
3. Recording accuracy and response Time: Brain Workshop automatically recorded each
participant's accuracy (percentage of correct responses) and response time per trial.
4. Saving results: Each session's data, including response accuracy and latency, was saved
automatically in a digital format for later analysis. If any issues arise with automatic
saving, participants should manually save their results and create a backup.

115